Improving Regenerating Mode Operation of MRAS-Based Induction Motor Speed Estimation Using the Multilayer Technique

Model Reference Adaptive Systems (MRASs) are one of the most commonly applied techniques to estimate induction motor speed. They ensure simple and algorithmic way to obtain all necessary variables to create a sensorless drive, or to provide the estimated speed signal for redundant system. However, one of the most popular MRAS-type estimators, MRAS-CC (C – Current estimator, C – Current flux simulator) is unstable in a large part of the regenerating mode operation. In this paper it is proposed to apply the multilayer observer theory to ensure the stability, while preserving large insensitivity to motor parameters. The proposed MRAS-ML (ML – MultiLayer) estimator includes two different MRAS-type estimators, unlike previous multilayer solutions with multiple identical observers, but with different parameters. Additionally, no auxiliary parameters need to be selected and it is not necessary to switch off any additional adaptation mechanism. It is shown that the MRAS-ML is stable in the whole regenerating mode operation using both simulation and experimental tests.